CPI (Consumer Price Index)

The percentage quoted is an annual inflation rate. To get to the stated figure, you have to divide the current month's CPI index figure by that of a year previously.

In message , N. Sloane writes

The 'headline' figure is quoted as month on the same month of the preceding year, so November 2012/November 2011. Not always a sensible way to do it, but usually better than just taking one month's change.

I'm with you at a point, but if the rate for November 2012 is 124.4 and the November 2011 figure is 121.2, do I just do a literal division ?

That would give an answer of 1.02644. Assume it is 2.644%, is the just rounded upwards to the nearest digit to give the rate of 2.7% ?

Thanks Neil.

It shouldn't be. What happens is that the CPI figures have been rounded to the nearest .1 and that gives rise to the occasional 0.1% error in the anual percentage rate.

This will give you less rounded figures than the source you are using.

Can you point me to the figures you have in mind please? The few times I have looked at the data in the ref tables they seemed all to be rounded as in the press releases etc rather than more precise data or computed from formulae - eg in

In message , Robin writes

You get one decimal place in the tables there (e.g. tables 1 or 2 in the reference you quoted) , which is more than the original poster was quoting. If you need more, you could contact them, but I don't know if they would want to publish any more detail.

Hi Sheila,

I have looked at the ONS figures and downloaded the Reference Tables. However, they are exactly the figures as the other source I was using.

So I'm still a bit confused how you can divide 124.4 by 121.2 and come out with a figure of 2.7%, because no matter how I try and round it, it always comes to

Am I missing some obscure rule about this ?

Neil.

If the actual figures were, say, 124.44 and 121.16, they would be rounded to 124.4 and 121.2. Dividing the more accurate figures gives a rounded 2.7%.

I am as puzzled as the OP who I see has already pointed out he quoted the same figures as in the tables.

FTAOD I expect it is indeed just a matter of the figures being rounded individually. And I am surprised there is not a prominent note to that effect.

In message , Robin writes

The obvious solution is on the ONS site, not here:

Contact

Philip Gooding

Prices

snipped-for-privacy@ons.gsi.gov.uk

Telephone: +44 (0) 1633 456900

I've contacted them. I'll be interested in their reply.

I rather doubt the ONS will tell me why you stated the tables give more decimal places than the OP quoted :)

Hi,

The ONS did get back to me about this.

Instead of calculating to one decimal place, i.e.

(124.4/121.2 - 1) x 100 = 2.6% to one decimal place. The actual calculation used by the ONS to calculate the annualised CPI is to three decimal places:

(124.378/121.153 - 1) x 100 = 2.7% to one decimal place.

The ONS will add you to a mailing list and send you the raw data on a monthly basis if you want to receive it.